Study And Application Of Algorithm For Polygon Aggregation In Geo-Cartographic Generalization

The digitalization of geo-maps and the broad establishment of geo-databases bring along a frequent need to produce small scale maps from large ones. As a result, automatic cartographic generalization has become the frontier and the main focus in this research field. Polygonal maps are a common data type in applications, such as geological maps, soil maps, and cadastral maps. As the map scale reduces, graphic conflicts raise between polygons. Operations have to been adopted to solve these conflicts, and aggregation is the most in common use. However, there are many shortcomings of current algorithms for polygon aggregation.In this thesis, an improved model for polygon aggregation based on constrained delaunay triangulation is put forward. The model mainly includes four parts: first clustering analysis of polygons, constrained delaunay triangulation, proximity analysis of polygons, and aggregation of polygons. First, a clustering analysis with map data is done in order to divide whole data into several parts. Second, a constrained delaunay triangulation is constructed for each part. Third, proximity analysis (second clustering analysis) is done to classify polygons with conflicts. At last, polygons with conflicts,βtriangle andγtriangle belong to them are aggregated.An automatic polygon aggregation system for the improved model is developed based on MapObjects and ArcSDE C-API in the environment of Visual C++ 6.0. It is then tried out in a geological map and a cadastral map. Applications show the advantage of this improved model for polygon aggregation.